15 species with 3 plants per species and 2 technical replicates of a vector (length of 27) metabolomic measurements on each plant. The metabolites can be split into sugars (3) and amino acids (24). Plant species can be categorized by Section (10 levels) and Pollinator species (5 levels). Also, pollinator species are not mutually exclusive, i.e. each plant species can be pollinated by more than 1 species. Each plant species flowers at day or night.
What affects metabolite expression? Section or pollinator species?
- PCA on 27 vector response (averaged over technical rep?). Unclear how to interpret.
- Analyze each metabolite separately. Average metabolite measurements across technical replicates. Fit cell means model, a mean for each species, and construct contrasts for pairwise comparisons among section and pairwise comparison among pollinator species.
- Analyze each metabolite separately. Average metabolite measurements across technical replicates and plants within a species. Fit a model with section and assess an F-test with pvalue for inclusion of section. Fit a model with pollinator species and assess an F-test with pvalue for inclusion of pollinator species.
With 15 species, making statements about differences among section and differences among pollinator species seems tenuous. In addition, these 15 species can be a large portion of the species within a section and therefore finite population corrections my be needed.
blog comments powered by Disqus